[PATCH kernel-4.19 1/3] jbd2: remove the out label in __jbd2_journal_remove_checkpoint()
From: Zhang Yi yi.zhang@huawei.com
hulk inclusion category: bugfix bugzilla: 50788 CVE: NA ---------------------------
The 'out' lable just return the 'ret' value and seems not required, so remove this label and switch to return appropriate value immediately. This patch also do some minor cleanup, no logical change.
Signed-off-by: Zhang Yi yi.zhang@huawei.com Reviewed-by: Jan Kara jack@suse.cz Reviewed-by: Zhang Xiaoxu zhangxiaoxu5@huawei.com Signed-off-by: Yang Yingliang yangyingliang@huawei.com --- fs/jbd2/checkpoint.c | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-)
diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c index 96bf33986d030..61de87fbf5ec3 100644 --- a/fs/jbd2/checkpoint.c +++ b/fs/jbd2/checkpoint.c @@ -562,13 +562,13 @@ int __jbd2_journal_remove_checkpoint(struct journal_head *jh) struct transaction_chp_stats_s *stats; transaction_t *transaction; journal_t *journal; - int ret = 0;
JBUFFER_TRACE(jh, "entry");
- if ((transaction = jh->b_cp_transaction) == NULL) { + transaction = jh->b_cp_transaction; + if (!transaction) { JBUFFER_TRACE(jh, "not on transaction"); - goto out; + return 0; } journal = transaction->t_journal;
@@ -577,9 +577,9 @@ int __jbd2_journal_remove_checkpoint(struct journal_head *jh) jh->b_cp_transaction = NULL; jbd2_journal_put_journal_head(jh);
- if (transaction->t_checkpoint_list != NULL || - transaction->t_checkpoint_io_list != NULL) - goto out; + /* Is this transaction empty? */ + if (transaction->t_checkpoint_list || transaction->t_checkpoint_io_list) + return 0;
/* * There is one special case to worry about: if we have just pulled the @@ -591,10 +591,12 @@ int __jbd2_journal_remove_checkpoint(struct journal_head *jh) * See the comment at the end of jbd2_journal_commit_transaction(). */ if (transaction->t_state != T_FINISHED) - goto out; + return 0;
- /* OK, that was the last buffer for the transaction: we can now - safely remove this transaction from the log */ + /* + * OK, that was the last buffer for the transaction, we can now + * safely remove this transaction from the log. + */ stats = &transaction->t_chp_stats; if (stats->cs_chp_time) stats->cs_chp_time = jbd2_time_diff(stats->cs_chp_time, @@ -604,9 +606,7 @@ int __jbd2_journal_remove_checkpoint(struct journal_head *jh)
__jbd2_journal_drop_transaction(journal, transaction); jbd2_journal_free_transaction(transaction); - ret = 1; -out: - return ret; + return 1; }
/*
From: Zhang Yi yi.zhang@huawei.com
hulk inclusion category: bugfix bugzilla: 50788 CVE: NA ---------------------------
Although we merged c044f3d8360 ("jbd2: abort journal if free a async write error metadata buffer"), there is a race between jbd2_journal_try_to_free_buffers() and jbd2_journal_destroy(), so the jbd2_log_do_checkpoint() may still fail to detect the buffer write io error flag which may lead to filesystem inconsistency.
jbd2_journal_try_to_free_buffers() ext4_put_super() jbd2_journal_destroy() __jbd2_journal_remove_checkpoint() detect buffer write error jbd2_log_do_checkpoint() jbd2_cleanup_journal_tail() <--- lead to inconsistency jbd2_journal_abort()
Fix this issue by introducing a new atomic flag which only have one JBD2_CHECKPOINT_IO_ERROR bit now, and set it in __jbd2_journal_remove_checkpoint() when freeing a checkpoint buffer which has write_io_error flag. Then jbd2_journal_destroy() will detect this mark and abort the journal to prevent updating log tail.
Signed-off-by: Zhang Yi yi.zhang@huawei.com Reviewed-by: Zhang Xiaoxu zhangxiaoxu5@huawei.com Signed-off-by: Yang Yingliang yangyingliang@huawei.com --- fs/jbd2/checkpoint.c | 12 ++++++++++++ fs/jbd2/journal.c | 14 ++++++++++++++ include/linux/jbd2.h | 11 +++++++++++ 3 files changed, 37 insertions(+)
diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c index 61de87fbf5ec3..b1af15ad36dcb 100644 --- a/fs/jbd2/checkpoint.c +++ b/fs/jbd2/checkpoint.c @@ -562,6 +562,7 @@ int __jbd2_journal_remove_checkpoint(struct journal_head *jh) struct transaction_chp_stats_s *stats; transaction_t *transaction; journal_t *journal; + struct buffer_head *bh = jh2bh(jh);
JBUFFER_TRACE(jh, "entry");
@@ -573,6 +574,17 @@ int __jbd2_journal_remove_checkpoint(struct journal_head *jh) journal = transaction->t_journal;
JBUFFER_TRACE(jh, "removing from transaction"); + + /* + * If we have failed to write the buffer out to disk, the filesystem + * may become inconsistent. We cannot abort the journal here since + * we hold j_list_lock and we have to careful about races with + * jbd2_journal_destroy(). So mark the writeback IO error in the + * journal here and we abort the journal later from a better context. + */ + if (buffer_write_io_error(bh)) + set_bit(JBD2_CHECKPOINT_IO_ERROR, &journal->j_atomic_flags); + __buffer_unlink(jh); jh->b_cp_transaction = NULL; jbd2_journal_put_journal_head(jh); diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index f1bc1c82c4802..89fad4c3e13cb 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -1429,6 +1429,10 @@ int jbd2_journal_update_sb_log_tail(journal_t *journal, tid_t tail_tid,
if (is_journal_aborted(journal)) return -EIO; + if (test_bit(JBD2_CHECKPOINT_IO_ERROR, &journal->j_atomic_flags)) { + jbd2_journal_abort(journal, -EIO); + return -EIO; + }
BUG_ON(!mutex_is_locked(&journal->j_checkpoint_mutex)); jbd_debug(1, "JBD2: updating superblock (start %lu, seq %u)\n", @@ -1784,6 +1788,16 @@ int jbd2_journal_destroy(journal_t *journal) J_ASSERT(journal->j_checkpoint_transactions == NULL); spin_unlock(&journal->j_list_lock);
+ /* + * OK, all checkpoint transactions have been checked, now check the + * write out io error flag and abort the journal if some buffer failed + * to write back to the original location, otherwise the filesystem + * may become inconsistent. + */ + if (!is_journal_aborted(journal) && + test_bit(JBD2_CHECKPOINT_IO_ERROR, &journal->j_atomic_flags)) + jbd2_journal_abort(journal, -EIO); + if (journal->j_sb_buffer) { if (!is_journal_aborted(journal)) { mutex_lock_io(&journal->j_checkpoint_mutex); diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index 667fce234ac90..5c0446f22bee1 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -780,6 +780,11 @@ struct journal_s */ unsigned long j_flags;
+ /** + * @j_atomic_flags: Atomic journaling state flags. + */ + unsigned long j_atomic_flags; + /** * @j_errno: * @@ -1276,6 +1281,12 @@ JBD2_FEATURE_INCOMPAT_FUNCS(csum3, CSUM_V3) * data write error in ordered * mode */
+/* + * Journal atomic flag definitions + */ +#define JBD2_CHECKPOINT_IO_ERROR 0x001 /* Detect io error while writing + * buffer back to disk */ + /* * Function declarations for the journaling transaction and buffer * management
From: Zhang Yi yi.zhang@huawei.com
hulk inclusion category: bugfix bugzilla: 50788 CVE: NA ---------------------------
Now that we can be sure the journal is aborted once a buffer has failed to be written back to disk, we can remove the journal abort logic in jbd2_journal_try_to_free_buffers() which was introduced in c044f3d8360d ("jbd2: abort journal if free a async write error metadata buffer"), because it may cost and propably is not safe.
Signed-off-by: Zhang Yi yi.zhang@huawei.com Reviewed-by: Jan Kara jack@suse.cz Reviewed-by: Zhang Xiaoxu zhangxiaoxu5@huawei.com Signed-off-by: Yang Yingliang yangyingliang@huawei.com --- fs/jbd2/transaction.c | 17 ----------------- 1 file changed, 17 deletions(-)
diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c index ffa6d3530f4bd..4055929a043cf 100644 --- a/fs/jbd2/transaction.c +++ b/fs/jbd2/transaction.c @@ -2085,7 +2085,6 @@ int jbd2_journal_try_to_free_buffers(journal_t *journal, { struct buffer_head *head; struct buffer_head *bh; - bool has_write_io_error = false; int ret = 0;
J_ASSERT(PageLocked(page)); @@ -2110,26 +2109,10 @@ int jbd2_journal_try_to_free_buffers(journal_t *journal, jbd_unlock_bh_state(bh); if (buffer_jbd(bh)) goto busy; - - /* - * If we free a metadata buffer which has been failed to - * write out, the jbd2 checkpoint procedure will not detect - * this failure and may lead to filesystem inconsistency - * after cleanup journal tail. - */ - if (buffer_write_io_error(bh)) { - pr_err("JBD2: Error while async write back metadata bh %llu.", - (unsigned long long)bh->b_blocknr); - has_write_io_error = true; - } } while ((bh = bh->b_this_page) != head);
ret = try_to_free_buffers(page); - busy: - if (has_write_io_error) - jbd2_journal_abort(journal, -EIO); - return ret; }
-
Yang Yingliang